native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated...

13
MARINE ECOLOGY PROGRESS SERIES Mar Ecol Prog Ser Vol. 558: 267–279, 2016 doi: 10.3354/meps11808 Published October 25 INTRODUCTION A number of classic studies in ecology have shown the importance of predators in shaping the structure of prey communities across diverse marine ecosystems. More specifically, substantial evidence supports the ability of resident fish predators to affect the size and structure of reef fish communi- ties via piscivory on post-settlement fish recruits (Hixon & Beets 1993, Stallings 2009). These studies have largely focused on the direct interactions be- tween predators and prey fish; however, predators can also affect the size and structure of prey com- munities via indirect interactions. For example, predators can shape prey communities when they alter the traits or behaviors of intermediate species, a phenomenon commonly known as behaviorally mediated indirect interactions (BMIIs; Strauss 1991). Interaction chains driven by changes in the behav- ior or traits of intermediate species are often at least as strong as density-driven effects and may in fact account for the majority of predator effects on food chains (see reviews by Werner & Peacor 2003 and Preisser et al. 2005). © R.D.E. and the USA Government 2016. Open Access under Creative Commons by Attribution Licence. Use, distribution and reproduction are unrestricted. Authors and original publication must be credited. Publisher: Inter-Research · www.int-res.com *Corresponding author: [email protected] Native grouper indirectly ameliorates the negative effects of invasive lionfish Robert D. Ellis 1,3, *, Meaghan E. Faletti 2 1 Florida State University, Department of Biological Science, Tallahassee, Florida 32306, USA 2 Florida Fish and Wildlife Conservation Commission, Division of Marine Fisheries Management, 2950 Executive Center Circle East, Tallahassee, Florida 32301, USA 3 Present address: Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, 100 8th Ave. SE, St. Petersburg, Florida 33701, USA ABSTRACT: Non-trophic interactions between Indo-Pacific lionfish Pterois volitans and P. miles and Atlantic and Caribbean reef fishes are not yet well understood. To determine the effects of potential competitive and behavioral interactions between native predators and invasive lionfish, we experimentally altered the presence of lionfish and red grouper Epinephelus morio in karst solution holes in Florida Bay, USA, and then tracked subsequent changes in the juvenile reef fish and motile macroinvertebrate communities for 6 wk. Relative to solution holes where we excluded both predators, mean juvenile reef fish abundance declined 83.7% in solution holes with a lionfish but increased by 154% in solution holes with a red grouper. There was no difference in juvenile reef fish abundance in solution holes with both lionfish and red grouper compared to holes where we excluded both predators. The composition of lionfish stomach contents shifted from mostly teleost fishes when lionfish were present in solution holes alone, to mostly crustaceans when in the presence of a red grouper. Concurrently, the abundance of 2 species of cleaner shrimp (Ancy- lomenes pedersoni and Periclimenes yucatanicus) decreased by 14.7% when lionfish were pres- ent but increased by 56.2% at holes where lionfish were excluded. We suggest that these results are due to altered lionfish predatory behavior in the presence of red grouper and highlight the importance of maintaining intact native predator communities for ameliorating the negative effects of the lionfish invasion. KEY WORDS: Lionfish · Red grouper · Interspecific interactions · Community ecology · Reef fish · Invasive species · Ecological impacts OPEN PEN ACCESS CCESS Contribution to the Theme Section ‘Invasion of Atlantic coastal ecosystems by Pacific lionfish’

Upload: others

Post on 20-May-2020

8 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 558: 267–279, 2016doi: 10.3354/meps11808

Published October 25

INTRODUCTION

A number of classic studies in ecology haveshown the importance of predators in shaping thestructure of prey communities across diverse marineeco systems. More specifically, substantial evidencesupports the ability of resident fish predators toaffect the size and structure of reef fish communi-ties via piscivory on post-settlement fish recruits(Hixon & Beets 1993, Stallings 2009). These studieshave large ly focused on the direct inter actions be -tween predators and prey fish; however, predators

can also affect the size and structure of prey com-munities via indirect interactions. For example,predators can shape prey communities when theyalter the traits or behaviors of inter mediate species,a phenomenon commonly known as behaviorallymediated indirect interactions (BMIIs; Strauss 1991).Interaction chains driven by changes in the behav-ior or traits of intermediate species are often atleast as strong as density-driven effects and may infact account for the majority of predator effects onfood chains (see reviews by Werner & Peacor 2003and Preisser et al. 2005).

© R.D.E. and the USA Government 2016. Open Access underCreative Commons by Attribution Licence. Use, distribution andreproduction are un restricted. Authors and original publicationmust be credited. Publisher: Inter-Research · www.int-res.com

*Corresponding author: [email protected]

Native grouper indirectly ameliorates the negativeeffects of invasive lionfish

Robert D. Ellis1,3,*, Meaghan E. Faletti2

1Florida State University, Department of Biological Science, Tallahassee, Florida 32306, USA2Florida Fish and Wildlife Conservation Commission, Division of Marine Fisheries Management,

2950 Executive Center Circle East, Tallahassee, Florida 32301, USA

3Present address: Florida Fish and Wildlife Conservation Commission, Fish and Wildlife Research Institute, 100 8th Ave. SE, St. Petersburg, Florida 33701, USA

ABSTRACT: Non-trophic interactions between Indo-Pacific lionfish Pterois volitans and P. milesand Atlantic and Caribbean reef fishes are not yet well understood. To determine the effects ofpotential competitive and behavioral interactions between native predators and invasive lionfish,we experimentally altered the presence of lionfish and red grouper Epinephelus morio in karstsolution holes in Florida Bay, USA, and then tracked subsequent changes in the juvenile reef fishand motile macroinvertebrate communities for 6 wk. Relative to solution holes where we excludedboth predators, mean juvenile reef fish abundance declined 83.7% in solution holes with a lionfishbut increased by 154% in solution holes with a red grouper. There was no difference in juvenilereef fish abundance in solution holes with both lionfish and red grouper compared to holes wherewe excluded both predators. The composition of lionfish stomach contents shifted from mostlyteleost fishes when lionfish were present in solution holes alone, to mostly crustaceans when in thepresence of a red grouper. Concurrently, the abundance of 2 species of cleaner shrimp (Ancy-lomenes pedersoni and Periclimenes yucatanicus) decreased by 14.7% when lionfish were pres-ent but increased by 56.2% at holes where lionfish were excluded. We suggest that these resultsare due to altered lionfish predatory behavior in the presence of red grouper and highlight theimportance of maintaining intact native predator communities for ameliorating the negativeeffects of the lionfish invasion.

KEY WORDS: Lionfish · Red grouper · Interspecific interactions · Community ecology · Reef fish ·Invasive species · Ecological impacts

OPENPEN ACCESSCCESS

Contribution to the Theme Section ‘Invasion of Atlantic coastal ecosystems by Pacific lionfish’

Page 2: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Mar Ecol Prog Ser 558: 267–279, 2016

Among the studies to explicitly test piscivory-dri-ven BMIIs, Stallings (2008) found a strong, positiveeffect of Nassau grouper Epinephelus striatus on theabundance of juvenile coral reef fishes in Bahamianpatch reefs that was driven by changes in the forag-ing behavior of small-bodied groupers. Pusack (2013)investigated a similar interaction in which he foundevidence for a BMII between Nassau grouper andjuvenile coral reef fishes mediated by the behaviorsof the lionfish Pterois volitans and P. miles, piscivoresnative to the Indo-Pacific and invasive in the westernAtlantic and Caribbean since at least 1985 (Schofield2009). Since about 2004, lionfish have experienced arapid range expansion throughout the Caribbean,Gulf of Mexico, and southeastern US Atlantic watersand have been present on reefs in the Florida Keys,USA, since at least 2009 (Ruttenberg et al. 2012). Asgeneralist predators, lionfish consume a diverse ar -ray of fishes and invertebrates (Valdez-Moreno et al.2012, Côté et al. 2013a), and some evidence suggestsan ontogenetic shift in diet where larger lionfish con-sume mostly fish (Morris & Akins 2009). On invadedcoral reefs, lionfish predation can reduce the abun-dance of native fishes by 80 to 94% and the biomassby up to 65% (Albins & Hixon 2008, Green et al.2012a, Albins 2013). By 2010 lionfish had invaded thehard bottom habitats of Florida Bay (see Fig. 1),which serves as an important nursery habitat formany fishes and invertebrates that move to nearbycoral reefs as adults (Fourqurean & Robblee 1999).Consequentially, predation by invasive lionfish inFlorida Bay on juvenile coral reef fishes may havecascading consequences for the health of nearbycoral reefs.

A common mesopredator resident in the hard bot-tom habitats of Florida Bay that will interact withinvasive lionfish is the red grouper Epinephelusmorio. In Florida Bay, red grouper are primarily asso-ciated with karst hard bottom features called solutionholes — pockmarked pits in the limestone formed bypast freshwater incursion — which they excavate byremoving sediment and detritus (Coleman et al.2010). Previous experiments conducted on the faunalcommunities associated with Florida Bay solutionholes showed that red grouper presence positivelyaffected the abundance and diversity of these com-munities and that the community-level effects weredriven by strong interactions with only a small number of individual species from the total speciespool (Ellis 2015). This group included some juvenilecoral-reef fishes, primarily small juvenile gruntsHaemulon spp., which were consistently among themost numerous fauna encountered in solution holes.

Red grouper consume primarily crustaceans andsome demersal fishes (Moe 1969, Weaver 1996) andare territorial, making aggressive displays that in -clude low frequency sound production and rapiddirect approaches to conspecifics and other residentsolution-hole fishes (Ellis 2015). These behaviors,combined with the fact that the red grouper is usuallythe largest individual animal encountered in solutionholes, may displace or disrupt predation by residentand transient predators around solution holes. Overtime, such behavioral interactions between redgrouper and other piscivores could result in differen-tial survival of post-settlement juvenile reef fishwhen compared to their survival in habitats withoutred grouper. This hypothetical BMII could be impor-tant in altering the predatory effects of lionfish thatinvade Florida Bay solution holes.

Several species of shrimp commonly found inFlorida Bay solution holes, including Periclimenesyucatanicus, Ancylomenes pedersoni, and Stenopushispidus, have been found in lionfish stomach con-tents (Morris et al. 2009, Faletti & Ellis 2014). At leastone of these species, A. pedersoni, is an experi -mentally verified cleaner that re moves ectoparasitesfrom reef fishes (Bunkley-Williams & Williams 1998,McCammon et al. 2010). It is yet unknown how thepresence of invasive lionfish may affect crustaceanbehaviors. Some shrimp species will change theirbehavior in the presence of finfish predators, oftenrelying on habitat features for protection (Ory & Thiel2013). Two common species of anemones in FloridaBay, Condylactis gigantea and Bartholomea annulata,have symbiotic relationships with cleaner shrimp andmay offer protection from predation for these species(Silbiger & Childress 2008, Briones-Fourzán et al.2012). Given that ectoparasites removed by cleanershrimp can have negative and even lethal conse-quences for parasitized fish (Artim et al. 2015), lion-fish predation on these species could representanother indirect negative effect on the native reeffishes in Florida Bay.

While other large-bodied groupers, including Nas-sau and tiger grouper Mycteroperca tigris, reportedlyprey on lionfish (Maljkovic et al. 2008), red grouperapparently do not (Morris 2009). Mesocosm experi-ments have shown little effect of native grouper pres-ence on lionfish behaviors (Morris 2009, Raymond etal. 2015). Furthermore, it has been widely debatedwhether mesopredators such as groupers actuallyfunction as biocontrol for invasive lionfish (see Mumbyet al. 2011 and subsequent responses by Hackerott etal. 2013 and Valdivia et al. 2014). However, Pusack(2013) reported that Nassau grouper appeared to

268

Page 3: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Ellis & Faletti: Grouper versus lionfish

reduce the effect of lionfish predation on native reeffish abundance. As congeners, Nassau and redgroupers are ex tremely similar in appearance andsize. Given these similarities, we were motivated toinvestigate if a similar behavioral interaction mayoccur between lionfish and red grouper. To do this,we set up an experiment to test the potential BMIIbetween red grouper and juvenile reef fish mediatedthrough lionfish in Florida Bay solution holes. Herewe present the results of an experiment designed toquantify (1) the effects of red grouper on solutionhole-associated juvenile reef fish abundance anddiversity; (2) the effects of lionfish on solution hole-associated juvenile reef fish abundance and diver-sity; and (3) the modification in lionfish effects onjuvenile reef fish abundance and diversity in thepresence of red grouper.

MATERIALS AND METHODS

Study site

Florida Bay is a large open embayment in southFlorida bordered by the Florida Keys, the Ever-glades, and the Gulf of Mexico (Fig. 1). The benthichabitat of Florida Bay is primarily seagrass, inter-spersed with areas of karst hard bottom that is usu-ally covered in a thin veneer of sediment and pock-marked with solution holes (Fourqurean et al. 2002).For this study, we used a set of 16 solution holes locatedin outer Florida Bay that were similar in terms of sizeand location (see Table S1 in the Supplement at

www. int-res.com/articles/ suppl/ m558 p267_ supp. pdf).So lu tion hole area, defined here as the product of the2 longest perpendicular measurements, ranged from1.69 to 6.99 m2 (mean ± SE: 4.20 ± 0.371 m2). Themaximum excavated depth of solution holes, definedhere as the deepest single measurement taken withinthe excavated area of the solution hole, ranged from26 to 77 cm (45.3 ± 3.01 cm).

Experimental design

To test the effects of lionfish on the abundance, di-versity, and community structure of juvenile reef fishassociated with solution holes in the presence and ab-sence of red grouper, we conducted a 6 wk field ex-periment where we manipulated the presence of bothred grouper and lionfish in solution holes. We first sur-veyed solution holes in early June 2013 to assess redgrouper and lionfish presence. All solution holes wereoccupied by a red grouper, but no lionfish were pres-ent at any of the solution holes at the start of the ex-periment. We randomly assigned the 16 red grouper-occupied solution holes to one of 4 treatment groups(n = 4): (1) no predators; (2) lionfish alone; (3) redgrouper alone; and (4) both lionfish and red grouper.Here we use the term ‘no predators’ to refer only tothe absence of red grouper and lionfish in this treat-ment; we did not manipulate the abundance of anyother native predators during the experiment. Weused the ‘no predator’ treatment to estimate the effectof all other native predators (e.g. black grouper, toad-fish, etc.) on solution hole communities. For the preda-

tor treatments, we used only a single li-onfish or single red grouper to matchthe typical densities of these predatorsas observed in Florida Bay solutionholes. Red grouper are generally soli-tary and often displayed territorial ag-gressive displays towards conspecifics(Ellis 2015). Lionfish densities in FloridaBay from 2010 to 2012 were similarlylow and, despite an increase in occur-rence over time, it was rare to en-counter more than a single individuallionfish in a solution hole. We assumedthat red grouper were unlikely to preyon lionfish but that red grouper pres-ence would disrupt lionfish predation.This assumption allowed us to ignoreany density-mediated indirect interac-tions (DMIIs) and just test for BMIIs be -tween red grouper and the suite of so-

269

Fig. 1. Approximate locations of solution hole sites in southwest Florida Bay,USA, used in this study. Open stars: sites where we conducted experimental

manipulations; closed stars: sites where we collected lionfish

Page 4: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Mar Ecol Prog Ser 558: 267–279, 2016

lution hole-associated juvenile coral reef fishes via li-onfish. For the purposes of estimating the BMII be-tween red grouper and juvenile reef fish via lionfish,the no predator treatment represents the base line‘control’ condition, the red grouper alone and lionfishalone treatments estimate the total effect of eachpredator separately on juvenile reef fish abundance,and the both predator treatment represents the‘threat’ condition (sensu Oku ya ma & Bolker 2007).

Red grouper present at the no predator and lionfishalone treatment holes were captured using hook-and-line, measured for total length (TL), tagged witha dart tag, and released at a vacant solution holelocated at least 5 km from the experimental studyarea. Previous experiments conducted in this systemshowed that relocating red grouper in this methodwas sufficient to prevent recolonization during theexperiment (Ellis 2015). We collected 8 lionfish fromhard bottom habitats located elsewhere in FloridaBay using monofilament hand nets, and then imme-diately transported them to the study area. Lionfishwere measured and then individually re leased ateach of the appropriate assigned lionfish alone orboth predator solution holes. Lionfish are known toexhibit high site fidelity (Jud & Layman 2012), so wedid not expect them to migrate between solutionholes during the study. The average size of lion -fish used during the 2013 experiments was 17.4 ±0.89 (SE) cm TL, which corresponds to an age of~4−5 mo (Potts et al. 2011).

Prior to predator manipulations, a team of 2 diverson SCUBA conducted a visual census of all reeffishes associated with all 16 solution holes, followingthe methods described by Hixon & Beets (1993). Afterslowly approaching the solution hole to a distance of~1 m from the edge, each diver slowly circled thehole while recording the number and identity of eachfish species. Divers first focused on the active plank-tivores hovering above the hole, then enumeratedand recorded any demersal and cryptic fishes andmacroinvertebrates found inside the hole using flash-lights to aid identification. Divers then summed theirrecorded abundances for each species and deter-mined the average number per species. Total speciesrichness was determined as the sum of all distinctspecies observed by both divers. Divers visually esti-mated the size of each fish to the nearest 1 cm (below10 cm TL) or to the nearest 5 cm (above 10 cm TL).Each survey lasted until all individuals were counted,or for a minimum of 5 min (mean census duration wasabout 12 min). The grunt species complex in FloridaBay contains at least 6 different species that are visu-ally indistinguishable at sizes <5 cm TL, so grunts

were identified to species when possible and individ-uals <5 cm were grouped together as ‘grunt recruits.’Divers also noted the identity and habitat associationof cleaner shrimp. Specifically, divers noted if shrimpwere found within 10 cm of either of the 2 anemonesfound in Florida Bay known to host cleaner shrimp(C. gigantea and B. annulata), or were found else-where in the solution hole not near anemones. Thisprotocol was repeated weekly for 6 wk (7 total sur-veys at each of the 16 solution holes). On average, ittook 2 d to survey all 16 experimental solution holes.

In addition to conducting weekly diver censuses,we checked the no predator and lionfish alone treat-ment holes once every 48 h for the duration of the ex-periment to ensure that no new red grouper orlionfish had moved onto these sites. New individualsencountered were captured, measured, tagged, andreleased at unoccupied sites as described above. Atthe end of the experiment all lionfish were collectedwith hand nets and euthanized with an overdoseof MS-222. We followed the methods described byGreen et al. (2012b) for all lionfish dissections: first,we recorded lionfish total length and dry blottedweight, then removed and weighed the stomach, en-tire alimentary canal, and all stomach contents. Werecorded the length and dry-blotted weight of allprey items before identifying prey to the lowest possi-ble taxonomic group using guides from Humann &Deloach (2002) for fishes and Abele & Kim (1986) forcrustaceans. Unrecognizable prey items were identi-fied using undigested hard parts (e.g. otoliths, skele-tons) whenever possible. Finally, we calculated theproportion of each prey group in the diet by number,size, weight, and frequency of occurrence.

Statistical analysis

We calculated a variety of community responsevariables to measure the effects of red grouper andlionfish separately and in concert on the native juve-nile reef fish populations associated with solutionholes: total abundance (N), Hill’s diversity numbersH0, H1, and H2, and Hill’s evenness (E). Hill’s num-bers provide a means of calculating commonly useddiversity indices using the single equation:

(1)

where pi is the relative proportion of the communitymade up by species i (Hill 1973). When evaluated forinteger values for a of 0, 1, and 2, Ha reduces to spe-cies richness, the antilog of the Shannon-Wiener

∑= ⎛⎝⎜

⎞⎠⎟

=

−H pa i

a

i

S a

1

11

270

Page 5: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Ellis & Faletti: Grouper versus lionfish

index, and the reciprocal of the Simpson’s index,respectfully. Generally, as a increases, the indexgives greater weight to more abundant species. Hill’sevenness, E, was calculated using the equation:

E2,1 = H2/H1 (2)

that Hill initially proposed because it does not in cludespecies richness (H0), and therefore is relatively insen-sitive to sample size (Hill 1997). Hill’s evenness con-verges to 1 when all species are equally abundant, sosmaller values indicate more uneven communities.

We used linear mixed-effects models (LMM) todraw inferences about the effects of predator treat-ment on the 5 community metrics (N, H0, H1, H2, andE), following the methods described by Albins (2013).The models included 2 categorical explanatory vari-ables, predator treatment and week, and a randomintercept for each solution hole. Time (week) was in -cluded as a categorical variable to eliminate any as-sumptions about the relationship between res ponsevariables and time. Solution hole identity was in-cluded as a random variable due to the repeatedmeasures design of the experiment that requiredmultiple observations of the same hole across time.Model selection was based on likelihood-ratio tests(LRT) performed on nested models to determine if in-cluding a treatment by the week interaction term im-proved the model fit. Visual examination of the modelresiduals suggested violations of the assumptions ofboth homogeneity of variance and inde pendence, sowe fit alternative models for each com munity re-sponse metric: one that incorporated heteroscedastic-ity among treatments, one that incorporated temporalautocorrelation among observations within solutionholes using the AR(1) autoregression model, and onethat incorporated both variance and autocorrelationstructures. Because the addition of the variance andautocorrelation structures caused the models to benon-nested, the resulting model fits were comparedwith Akaike’s information criterion (AIC) to deter -mine the optimal model for hypothesis testing. t-testswere used to evaluate differences between all treat-ment contrasts for each of the 5 community responsemetrics at the end of the experiment, when the best-fit model from the LMM analysis indicated that pred-ator treatment had a significant effect.

To evaluate the effects of red grouper and lionfishon the structure of the juvenile reef fish communities,we performed ordinations with non-metric multidi-mensional scaling (NMDS; McCune & Grace 2002).For all ordinations, recruit abundances were square-root transformed and standardized using the Wiscon-sin double standardization, where abundances were

first standardized by species maxima and then by thesample total. We calculated Bray-Curtis distances forthe ordinations and for hypothesis testing of recruitcommunity structure and then tested for differencesin the composition of recruit communities with per-mutational multivariate analysis of variance (PERM-ANOVA; Anderson 2001). Recruit abundances werenot transformed or standardized, and all analyseswere run with 1000 unconstrained permutations. Sta-tistical analyses were conducted in the R softwareenvironment (R Core Team 2014) using the ‘lme4’package for the LMMs (Bates et al. 2015), and the‘MASS’ and ‘vegan’ packages for the NMDS andPERMANOVA analysis (Venables & Ripley 2002,Oksanen et al. 2011).

We estimated effect sizes for the direct effects oflionfish, the total indirect effects of red grouper, andthe effects of both predators together on juvenile reeffish recruit abundance using a ratio-based approach(Trussell et al. 2006, Hughes et al. 2012). The directeffect (DE) of lionfish (LF) on recruit abundance, R,was calculated with the ratio of recruit abundance inthe lionfish alone treatment to the mean recruitabundance in the no predator (NP) treatment:

DELF = (RLF/RNP) − 1 (3)

Similarly, the indirect effect (IE) of red grouper onrecruit abundance was calculated with the ratio ofrecruit abundance in the red grouper alone treat-ment (RG) to the mean recruit abundance in the nopredator treatment:

IERG = (RRG/RNP) − 1 (4)

Finally, the BMII of red grouper on recruit abun-dance via lionfish was calculated with the ratio of therecruit abundance with both predators (BP) (i.e. theeffect of both predators or the lionfish effect in thepresence of predator cues) to the mean recruit abun-dance in the lionfish alone treatment (followingOkuyama & Bolker 2007):

BMII = (RBP/RLF) − 1 (5)

The numerators for all ratios were provided by allreplicates of the given treatment, whereas the de -nominator was the mean recruit abundance at theend of the experiment for the given treatment. Weestimated means and 95% confidence intervals forall effect sizes by bootstrapping 1000 times withreplacement to account for the low number of exper-imental replicates available for each of the treat-ments. This approach is similar to the methods usedby Paine (1992) to quantify interaction strength.

In an earlier analysis, we found a relatively high

271

Page 6: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Mar Ecol Prog Ser 558: 267–279, 2016

proportion of decapod crustaceans in thestomach contents of lionfish collected fromFlorida Bay including some species knownto provide cleaning services (Faletti & Ellis2014), so we were motivated to test forpredator treatment effects on shrimp abun-dance using the LMM framework describedabove. When specific treatment effectswere not found, we grouped treat ments bylionfish presence — lionfish present (eitheralone or with a red grouper) and lionfishabsent (no predators or red grouperalone) — and tested for differences in shrimpabundance between the start and end ofthe experiment using paired-sample t-tests. We tested for differences in cleanershrimp habitat association between thestart and end of the experiment and for dif-ferences in the fish to invertebrate ratio inlionfish stomach contents in the presenceand absence of red grouper using a 2-tailedFisher’s exact test. We set the level of statis-tical significance for all tests at α = 0.05 andconsidered results to be marginally signifi-cant when 0.05 < p ≤ 0.10. Except whennoted otherwise, all results are presentedas means ± SE.

RESULTS

Predator effects

After 6 wk the abundance of juvenilereef fish re cruits was greatest at solutionholes with red grouper alone (111 ± 35.4),followed by holes with both predators (50.8± 15.2), and holes with neither predator(43.8 ± 17.2; Fig. 2). Of the 4 solution holesof the lion fish alone treatment, 3 had zerorecruits present after 6 wk, while at thefourth we counted 33 recruits (8.25 ± 8.25).On average, the abundance of juvenilereef fish recruits at solution holes with bothpre dators was significantly greater com-pared to the lionfish alone treatment andless than at the red grouper alone holes,but was not different from the no predatorholes (Table 1). The re sults of LMM analy-sis supported including predator treatmenteither as a main effect or as an interactionwith time for all metrics (see Table S2 inthe Supplement). The preferred variance

272

Response Treatment Estimate Contrast p

Abundance NP vs. LF** 0.037NP 43.8 ± 17.2 NP vs. RG** 0.050LF 8.25 ± 8.25 NP vs. BP 0.413RG 111 ± 35.4 LF vs. RG** 0.023BP 50.8 ± 15.2 LF vs. BP** 0.027

RG vs. BP* 0.059

Species richness (H0) NP vs. LF** 0.010NP 4.50 ± 0.87 NP vs. RG 0.148LF 1.00 ± 1.00 NP vs. BP 0.296RG 3.25 ± 0.63 LF vs. RG** 0.014BP 5.75 ± 0.63 LF vs. BP** <0.001

RG vs. BP** 0.031

Shannon diversity (H1) NP vs. LF* 0.059NP 2.78 ± 0.17 NP vs. RG 0.171LF 0.86 ± 0.86 NP vs. BP 0.448RG 2.30 ± 0.41 LF vs. RG 0.104BP 3.24 ± 0.71 LF vs. BP** 0.033

RG vs. BP 0.208

Simpson diversity (H2) NP vs. LF* 0.077NP 2.25 ± 0.08 NP vs. RG 0.722LF 0.78 ± 0.78 NP vs. BP 0.470RG 2.09 ± 0.41 LF vs. RG* 0.097BP 2.55 ± 0.35 LF vs. BP* 0.053

RG vs. BP 0.427

Evenness (H2/H1) NP vs. LF* 0.080NP 0.82 ± 0.03 NP vs. RG* 0.083LF 0.22 ± 0.22 NP vs. BP 0.663RG 0.90 ± 0.02 LF vs. RG* 0.058BP 0.80 ± 0.03 LF vs. BP* 0.086

RG vs. BP** 0.042

Table 1. Estimated mean ± SE of juvenile reef fish recruit communityabundance, species richness, diversity, and evenness. p-values frompairwise t-tests performed on each of the 6 a priori contrasts based onthe 5 community response variables at the end of the 6 wk experiment;n = 4 per predator treatment; *p ≤ 0.1, **p ≤ 0.05. NP: no predator; LF:

lionfish alone; RG: red grouper alone; BP: both predators

C

BB

A

0

25

50

75

100

125

0

150

1 2 3 4 5 6

Rec

ruit

abun

dan

ce

Time (week)

Red grouper aloneBoth predatorsNo predatorsLionfish alone

Fig. 2. Abundance (mean ± SE) of juvenile reef fish recruits at solutionholes in Florida Bay from 4 experimental predator treatments (n = 4 perpredator treatment). Letters at the far right: results of pairwise compar-isons among treatments at the final census; matching letters indicatep > 0.05 based on pairwise t-tests performed on the best-fit linear mixed-

effects model

Page 7: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Ellis & Faletti: Grouper versus lionfish

structure and in clusion of temporal autocorrelationdiffered among the metrics tested, i.e. allowing vari-ance to vary by treatment improved the models forabundance and evenness while including temporalautocorrelation improved the models for abundanceand species richness.

Red grouper and lionfish had varying effects on theother community metrics when compared to the nopredator holes. Red grouper alone had a marginallysignificant positive effect on the evenness of juvenilereef fish recruits compared to the no predator treat-ment (red grouper alone = 0.900 ± 0.021; no preda-tors = 0.816 ± 0.032; p = 0.083), but no other compar-isons were significant (see Table 1). However,lionfish alone had a significant negative effect on all5 of the community metrics tested compared to the nopredator holes (Fig. 3). The species richness, Simp-son’s diversity, and evenness of juvenile reef fishcommunities were all significantly greater with redgrouper compared to holes with lionfish alone. Ingeneral, communities were more species rich, diverse,and even with both predators compared to holes withlionfish alone (Fig. 3).

Compared to the no predator treatment, communi-ties with both predators were not significantly differ-ent in terms of any of the community response met-rics analyzed, including recruit abundance. However,we found a marginally significant difference be -tween communities in the presence of both predators(5.75 ± 0.63 recruit species) compared to those withred grouper alone (3.25 ± 0.63 recruit species; p =0.059), but these communities were significantly lesseven with both predators (0.80 ± 0.03) than they werewith red grouper alone (0.90 ± 0.02; p = 0.042).

The structure of recruit communities varied greatlyamong predator treatments. During the 6 wk experi-ment, we encountered 14 unique species of juvenilereef fishes. However, we counted only 9 of these spe-cies during the final survey at the end of the experi-ment (Table 2). Of this group, the grunts, specificallywhite and French grunts (Haemulon plumerii and H.

273

Predator treatment

A

C

A,B

B

0

2

4

6

RG LF NP BP

a

Sp

ecie

s ric

hnes

s

A,B

A

B

B

0.0

1.0

2.0

3.0

4.0

RG LF NP BP

b

Sha

nnon

div

ersi

ty

A

B

A

A

0.0

1.0

2.0

3.0

RG LF NP BP

c

Sim

pso

n d

iver

sity

A

B

C C

0.0

0.2

0.4

0.6

0.8

1.0

RG LF NP BP

d

Hill

’s e

venn

ess

Fig. 3. (a) Species richness, (b) Shannon and (c) Simpson di-versity, and (d) Hill’s evenness of juvenile reef fish recruitcommunities after 6 wk with red grouper alone (RG; d), lion-fish alone (LF; M), no predator (NP; s), or both predators (BP;R); n = 4 per predator treatment. Different letters indicatesignificant differences between groups after 6 wk based onpairwise t-tests performed on the best-fit linear mixed model

Species Common name Control RG effect LF effect Combined effect BMII NNP (NRG−NNP) (NLF−NNP) (NBP−NNP) (NBP−NLF)

Haemulon plumierii White grunt 17.3 47.8 −13.5 8.5 22.0Haemulon flavolineatum French grunt 11.3 14.5 −9.0 2.75 11.8Haemulon spp. Juvenile grunts 11 7.75 −9.5 −4.5 5Anisotremus virginicus Porkfish 0.25 0.25 −0.25 0.5 0.75Pomacanthus arcuatus Gray angelfish 0.0 0.0 0.0 0.75 0.75Holacanthus ciliaris Queen angelfish 0.25 −0.25 −0.25 0.0 0.25Haemulon parra Sailors choice 1.0 −0.5 −1.0 0.0 1.0Lutjanus synagris Lane snapper 0.5 −0.5 −0.5 −0.5 0.0Pareques acuminatus Highhat 2 −1.25 −1.25 −1.0 0.25

Total 43.5 67.8 −35.3 6.5 41.8% change − 156 −81.1 14.9 96.1

Table 2. Treatment effects, in terms of the relative difference in mean abundance of each of the 9 species encountered duringthe final (Week 6) survey and the full juvenile reef fish recruit community, for red grouper alone (RG), lionfish alone (LF), bothpredators together (BP), and the BMII estimate. Control values presented are the mean abundance of each species in the

no predator (NP) treatment

Page 8: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Mar Ecol Prog Ser 558: 267–279, 2016

flavolineatum, respectively) were the most numer-ous. NMDS ordination of recruit communities show -ed clear separation of communities with lionfish(Fig. 4). However, because 3 of the 4 lionfish alonesolution holes had zero recruits at the end of theexperiment, the 3 zero-abundance points overlaid ontop of each other (see point at [−1.33, 0.02] in Fig. 4),so the ellipse representing the standard deviation ofthe lionfish alone treatment communities collapsedto a line. There was significant overlap in the redgrouper alone and no predator communities, whilecommunities with both predators appeared to be sep-arate from all other groups. PERMANOVA resultsindicated that predator treatment had a significanteffect on community structure (pseudo-F3,15 = 2.27,p = 0.029), supporting the separation of communitiesby treatment visualized in the NMDS.

Overall lionfish alone reduced juvenile reef fishrecruit abundance by 81.1% compared to the nopredator treatment (Table 2). The bootstrapped esti-mate of the direct effect of lionfish on recruit abun-dance, calculated as the ratio of recruit abundancewith lionfish to recruit abundance with no predators,was −0.802 (range: −1.00 to −0.434). The estimated

indirect effect of red grouper alone on recruitabundance was 1.546 (range: 0.206–2.77), inline with the 156% increase in re cruit abun-dance observed during the experiment. Theestimated effect size of the BMII between redgrouper and juvenile reef fish recruits vialionfish according to the ratio-based methodwas estimated as 5.18 (range: 2.15–8.73) orabout 5 times the expected recruit abun-dance with lionfish alone.

After 6 wk, we found no effect of predatortreatment on cleaner shrimp abundance. TheLRT did not support including predator treat-ment in the final model for shrimp abun-dance (likelihood-ratio9,8 = 0.77; p = 0.38).However, the mean abundance of cleanershrimp declined by 14.7% in solution holeswith lionfish between the start and end of theexperiment, irrespective of the presence orabsence of red grouper. Conversely, meancleaner shrimp abundance in solution holeswithout lionfish increased by 56.2%, a mar-ginally significant effect (t = −2.23; p = 0.052;Fig. 5). We also observed a significant shift incleaner shrimp association with anemones insolution holes with lionfish, where 42.3% ofcleaner shrimp were found within 10 cm ofan anemone at the start of the experimentand 61.2% were within 10 cm of an anemone

at the end of the experiment (p = 0.011). In solutionholes without lionfish, cleaner shrimp as so ciationwith anemones did not significantly change (p = 0.89;Fig. 6).

274

p = 0.54

0

5

p = 0.052

10

15

20

25

30

Lionfish No lionfish

StartEnd

Cle

aner

shr

imp

ab

und

ance

Fig. 4. Non-metric multidimensional scaling (NMDS) ordination ofjuvenile reef fish communities (abundance by species) associatedwith experimental solution holes at the end of the 6 wk experimentconducted in Florida Bay during June−July, 2013. Ellipses show theSD of all points for each predator treatment group; n = 4 per predatortreatment. No ellipse for the lionfish alone treatment as 3 of the 4 juvenile fish communities were non-existent (n = 0) at the end of theexperiment; these sites are represented by the single point located

at (−1.33, 0.0217)

Fig. 5. Comparison between cleaner shrimp abundance (mean± SE) at the start and end of the experiment at solution holeswith (lionfish alone + both predator treatments) and without(no predator + red grouper alone treatments) lionfish. p-values reported above bars were based on paired t-testscomparing shrimp abundance at the start and end of the

experiment for each group

Page 9: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Ellis & Faletti: Grouper versus lionfish

Lionfish diet analysis

Between June and August 2013 we analyzed thestomach contents of 29 lionfish captured at solutionholes with (n = 13) and without (n = 16) red grouper.Overall, fish made up 51% of stomach contents bynumber and crustaceans made up 35.4%, with the re-maining 13.5% of prey items unidentifiable (seeTable S3 in the Supplement). Mean gut contentweight was 1.92 ± 0.56 g; only one stomach was empty.Palaemonid shrimps were the largest consumed fam-ily by number (19.8%), while collectively cleanershrimps (e.g. Lysmata spp., Ancylomenes pedersoniand Periclimenes yucatanicus) made up 30.1% oflion fish stomach contents. By weight, fish prey com-prised 87.1% of lionfish stomach contents and crus-taceans comprised 9.68%; 3.23% of prey items byweight were unidentifiable. Within teleost preygroups, Lutjanids made up the largest percentage byweight (34.2%), followed by unidentifiable teleosts(22.3%), grunts (genus Haemulon; 16.4%) and gobies(family Gobiidae; 9.43%). We documen ted a shift inlionfish diet in the presence of red grouper from a pri-marily piscivorous diet when the lionfish were in solu-tion holes alone to a crustacean-based one in holeswhen both lionfish and red grouper were present to-gether (p = 0.028; Fig. 7). When alone, lionfish con-sumed 78.4% fish by number but fish prey made upjust 43.4% of lionfish stomach contents when a redgrouper was also present at the time of capture.

DISCUSSION

The results of our experiment and dietanalysis suggest that red grouper enhancethe abundance of juvenile reef fishes thatrecruit to solution holes in Florida Bay andsupport the hypothesis that this effect occursvia changes in piscivore behavior in solutionholes. Juvenile reef fishes benefited from thepresence of the relatively large, territorial,habitat manipulating red grouper, while,conversely, recruit abundance was signifi-cantly depleted with lionfish. Lionfish areextremely efficient predators on Caribbeanreef fishes, and reef fish populations on coralreefs and hard bottom habitats invaded bylionfish have suffered significant declinesshortly following invasion (Albins & Hixon2008, Green et al. 2012a). The results pre-sented here confirm that the negative effectsof lionfish on native reef fish populationsobserved elsewhere in the invaded range

275

p = 0.011

p = 0.89

0.00

0.20

0.40

0.60

0.80

Start End Start End

Lionfish

Pro

por

tion

asso

ciat

ed

Hole

No lionfish

Anemone

Fig. 6. Proportion of cleaner shrimp encountered on or within 10 cmof an anemone (‘Anemone’) or more than 10 cm from an anemone(‘Hole’) at the start versus the end of the experiment. Predator treat-ments were grouped by lionfish presence (lionfish alone + both pre -dators treatments) or absence (no predator + red grouper alonetreatments) in solution holes; n = 8 for each group. p-values reportedabove bars are based on Fisher’s exact test results comparingshrimp− anemone association at the start and end of the experiment

for each group

0

10

20

30

40

50

60

70

80

90

100

Grouper present

Die

t co

mp

ositi

on (%

)

Grouper absent

GobiidaeHaemulonLutjanidaeUID teleostPalaemonidaeLysmataHippolytidaeUID shrimpUID crustacean

Fig. 7. Diet composition (% by number) of lionfish collectedfrom solution holes in Florida Bay with (‘Grouper present’)and without (‘Grouper absent’) a red grouper at the time of

collection. UID: unidentifiable

Page 10: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Mar Ecol Prog Ser 558: 267–279, 2016

also hold true for Florida Bay hard bottom fish communities.

Lionfish and red grouper had very different effectson the diversity of juvenile reef fishes in Florida Baysolution holes. In general, solution holes with redgrouper alone were similar to those with neitherpredator in terms of species richness and diversity.Meanwhile, similar comparisons are difficult to makefor lionfish as they reduced recruitment to zero at 3 ofthe 4 lionfish-only treatment holes. While all 4 of thecommunity diversity metrics were depressed in thelionfish alone treatment compared to the others,making specific conclusions about predator-driveneffects on diversity are problematic when the preda-tor leaves no prey at all. Certainly, this is a strongqualitative result if not a quantitative one: lionfishreduce the abundance and diversity of native Carib-bean reef fishes, apparently through indiscriminatepiscivory. Our results parallel those found in otherparts of the invaded range including reductions innative reef fish recruitment (Albins & Hixon 2008),reef fish abundance (Green et al. 2012a, Côté et al.2013b), and species richness (Albins 2015).

We found no differences in recruit abundance insolution holes with both predators compared to theother treatments, and solution holes with both preda-tors had comparable species richness and diversitycompared to the red grouper alone and no predatortreatments (see Table 1). These results generally sup-port the BMII hypothesis: when red grouper werepresent in solution holes with lionfish, the associatedcommunities of juvenile reef fishes, although de pres -sed in abundance, were still similar to those in solu-tion holes without red grouper or without eitherpredator. Native fish communities appeared to bene-fit from the presence of red grouper compared towhen the exotic piscivorous lionfish was presentalone.

Our reevaluation of the diet composition of lionfishfound in Florida Bay solution holes revealed thatteleosts made up a smaller portion of the diet bynumber in Florida Bay compared to other diet studiesdone in the Bahamas and North Carolina (Morris etal. 2009, Munoz et al. 2011). We also found that gob-ies had the highest index of relative importance (IRI)of all prey groups, supporting reports elsewhere thatgobies may be especially vulnerable to lionfish pre-dation (Morris & Akins 2009, Albins & Lyons 2012).The shift we observed in diet from teleost-dominatedto crustacean-dominated appeared to be driven bythe presence of red grouper. This result further supports our BMII hypothesis that the disruptivepresence of red grouper cause lionfish to alter their

predation behavior to consume more benthic crusta -ceans that may be easier prey to consume comparedto juvenile reef fish. Further investigation is neededto decipher the specific mechanisms driving thisinteraction. However, if lionfish alter their diets totarget certain species such as cleaner shrimp, thiscould ultimately lead to a loss of the ecosystem serv-ices provided by those species. Larval settlement pat-terns for these 2 species in Florida Bay are un known,though some other species of shrimps exhibit peaksof post larval settlement in the Middle Keys range ofFlorida Bay during the summer months (Cria les et al.2006). A summer influx of post larval shrimps couldexplain the increase in shrimp abundance we ob -served in the absence of lionfish through out thestudy period.

Red grouper do not compete with lionfish for prey;although both will consume crustaceans and demer-sal fishes, the diets of lionfish collected from FloridaBay do not overlap with red grouper diets. However,we suspect that lionfish and red grouper may com-pete for space in solution holes. The exact nature ofthe lionfish response to red grouper, via modificationof some specific behaviors by the lionfish or a moregenerally disruptive effect of red grouper presence,remains untested. The few studies that have ex -plicitly investigated behavioral inter actions betweennative Atlantic reef fishes and lionfish have foundthat lionfish generally ignore potential predatorswhile native fish actively avoid the lionfish. Onestudy of competitive shelter use between lionfish andNassau grouper in experimental mesocosms foundthat Nassau grouper avoided lionfish even when theywere much larger than the lionfish, but lionfishdid not change their use of shelter even when thegrouper was much larger (Raymond et al. 2015). During an experimental feeding trial, Morris (2009)reported that red grouper moved away from lionfishwhen approached. In the present study, we did notobserve avoidance of lionfish by red grouper, or viceversa. However, the duration of such observationswas limited to the time we spent conducting commu-nity censuses, and we did not quantify avoidancebehaviors by either fish. Nevertheless, our resultsand those reported by Pusack (2013) suggest thatinvasive lionfish alter their feeding behaviors in thepresence of larger native groupers.

The link between recruitment, post-settlementmortality, and adult population size for reef fish iscomplex. However, Shulman & Ogden (1987) foundthat changes in immediate post-settlement survivalof French grunts was a more important factor in reg-ulating the ultimate abundance of adult grunts on

276

Page 11: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Ellis & Faletti: Grouper versus lionfish

coral reefs than to changes in recruitment. We testedthe hypothesis that enhanced reef fish abundance atred grouper occupied solution holes results in partfrom disruptive behavioral interactions with pisci-vores that lead to enhanced survival of juvenile reeffishes. Im portantly, we did not observe any predationon the transplanted lionfish during the experiment,despite reports elsewhere that groupers may act aspredators of invasive lionfish (Maljkovic et al. 2008,Mumby et al. 2011). There are increasing controlefforts across the invaded range, including attemptsby spearfishers to ‘teach’ native groupers and sharksto feed on lionfish that have been largely unsuccess-ful and dange rous for divers (Diller et al. 2014). Ulti-mately, it may be that intact native fish communitiesthat include native mesopredators are the best wayto ameliorate the worst-case effects of lionfish bycompeting with and altering their behavior (Albins& Hixon 2013). Some recent evidence suggests thatCaribbean reefs with relatively high native predatordensity can maintain unchanged prey populationsdespite being invaded by lionfish (Elise et al. 2015).Other studies have found that lionfish-induced re -ductions in the density of native fishes <10 cm TL didnot translate to larger prey (10–20 cm TL; Albins2015). Lionfish, then, may represent an en hancedgauntlet that juvenile reef fish must pass through,but not necessarily an impenetrable one.

The successful invasion of exotic species is thoughtto be more likely in human-altered ecosystems (Sax& Brown 2000). As humans have reduced the bio-mass of native mesopredators on coral reefs and hardbottom habitats throughout the Caribbean, there arefewer competitive and behavioral interactions thatlimit both the population size and predatory effects oflionfish in its native range. The experiment presentedhere supports the mesopredator release hypothesisby showing how the presence of a native predator,the red grouper, can ameliorate the negative effectsof lionfish predation on native reef fish communities.This study sheds some light on the community-leveleffects of both native and invasive predators, an inte-gral part of expanding fisheries management fromsingle-species stock assessments to ecosystem-basedfisheries management (Pikitch et al. 2004). Here weshow how an important fishery species, the redgrouper, has complex direct and indirect interactionswith the other species that colonize grouper-exca-vated solution holes in Florida Bay. Some of theseinteractions may have population-level effects onspecies that support fisheries and provide importantecosystem services, services that are lost or reducedby the presence of lionfish (Johnston et al. 2015).

Albins & Hixon’s (2013) description of a ‘worst-casescenario’ for lionfish in the western Atlantic high-lights the need for intact predator communities toameliorate the effects of the lionfish invasion. Ourstudy provides some of the first experi mental evi-dence of this effect and begins to shed light on themechanisms by which native predators may lessenthe negative effects of this exotic invader.

Acknowledgements. This project was conducted with per-missions from the Florida Keys National Marine Sanctuary(permit #FKNMS-2012-078), the Florida Fish and WildlifeConservation Commission (SAL #11-1330-SR), and the FSUInstitutional Animal Care and Use Committee (IACUC Pro-tocol #1106). Financial support was provided by grants fromthe PADI Foundation, Florida Sea Grant, and the Guy Har-vey Ocean Foundation to R.D.E. We thank C. Ellis, C. Mali-nowski, T. Richards, A. Schmidt, and M. Wilkerson for theirassistance in the field. The FSU Coastal and Marine Labora-tory provided valuable logistical support. Conversationswith F. Coleman, M. Hixon, C. Koenig, and J. Wulff con-tributed significantly to the design of this work, while themanuscript benefited from comments by T. Miller, E. DuVal,M. Huettel, B. Inouye, and 3 anonymous reviewers. Finally,we thank the Gulf and Caribbean Fisheries Institute (GCFI)for their support of lionfish research and for a Travel Awardto R.D.E. that enabled our participation at the 68th GCFI inPanama.

LITERATURE CITED

Abele LG, Kim W (1986) An illustrated guide to the marinedecapod crustaceans of Florida. Fla Dept Environ RegTech Ser, Vol 8. Florida State University, Tallahassee, FL

Albins MA (2013) Effects of invasive Pacific red lionfishPterois volitans versus a native predator on Bahamiancoral-reef fish communities. Biol Invasions 15: 29−43

Albins MA (2015) Invasive Pacific lionfish Pterois volitansreduce abundance and species richness of native Baha -mian coral-reef fishes. Mar Ecol Prog Ser 522: 231−243

Albins MA, Hixon MA (2008) Invasive Indo-Pacific lionfishPterois volitans reduce recruitment of Atlantic coral-reeffishes. Mar Ecol Prog Ser 367: 233−238

Albins MA, Hixon MA (2013) Worst case scenario: potentiallong-term effects of invasive predatory lionfish (Pteroisvolitans) on Atlantic and Caribbean coral-reef communi-ties. Environ Biol Fishes 96: 1151−1157

Albins MA, Lyons PJ (2012) Invasive red lionfish Pteroisvolitans blow directed jets of water at prey fish. Mar EcolProg Ser 448: 1−5

Anderson MJ (2001) A new method for non-parametricmulti variate analysis of variance. Austral Ecol 26: 32−46

Artim JM, Sellers JC, Sikkel PC (2015) Micropredation bygnathiid isopods on settlement-stage reef fish in the east-ern Caribbean Sea. Bull Mar Sci 91: 479−487

Bates D, Maechler M, Bolker B, Walker S (2015) Fitting linear mixed-effects models using lme4. J Stat Soft 67:1–48

Briones-Fourzán P, Pérez-Ortiz M, Negrete-Soto F, Bar-radas-Ortiz C, Lozano-Álvarez E (2012) Ecological traitsof Caribbean sea anemones and symbiotic crustaceans.Mar Ecol Prog Ser 470: 55−68

277

Page 12: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Mar Ecol Prog Ser 558: 267–279, 2016

Bunkley-Williams L, Williams EH (1998) Ability of Pedersoncleaner shrimp to remove juveniles of the parasiticcymothoid isopod, Anilocra haemuli, from the host.Crusta ceana 71: 862−869

Coleman FC, Koenig CC, Scanlon KM, Heppell S, HeppellS, Miller MW (2010) Benthic habitat modificationthrough excavation by red grouper, Epinephelus morio,in the northeastern Gulf of Mexico. The Open Fish Sci J3: 1−15

Côté IM, Green SJ, Morris JA Jr, Akins JL, Steinke D(2013a) Diet richness of invasive Indo-Pacific lionfishrevealed by DNA barcoding. Mar Ecol Prog Ser 472: 249−256

Côté IM, Green SJ, Hixon MA (2013b) Predatory fishinvaders: insights from Indo-Pacific lionfish in the west-ern Atlantic and Caribbean. Biol Conserv 164: 50−61

Criales MM, Wang JD, Browder JA, Robblee MB, JacksonTL, Hittle C (2006) Variability in supply and cross-shelftransport of pink shrimp (Farfantepenaeus duorarum)postlarvae into western Florida Bay. Fish Bull (Wash DC)104: 60−74

Diller JL, Frazer TK, Jacoby CA (2014) Coping with the lion-fish invasion: evidence that naïve, native predators canlearn to help. J Exp Mar Biol Ecol 455: 45−49

Elise S, Urbina-Barreto I, Boadas-Gil H, Galindo-Vivas M,Kulbicki M (2015) No detectable effect of lionfish (Pteroisvolitans and P. miles) invasion on a healthy reef fishassemblage in Archipelago Los Roques National Park,Venezuela. Mar Biol 162: 319−330

Ellis RD (2015) Ecological effects of red grouper (Epineph-elus morio) in Florida Bay. PhD dissertation, Florida StateUniversity, Tallahassee, FL

Faletti ME, Ellis RD (2014) Novel predator, novel habitat: adiet analysis and experimental test of the ecologicaleffects of invasive lionfish in Florida Bay. Proc GulfCaribb Fish Inst 66: 217−221

Fourqurean JW, Robblee MB (1999) Florida Bay: a history ofrecent ecological changes. Estuaries 22: 345−357

Fourqurean JW, Durako MJ, Hall MO, Hefty LN (2002) Sea-grass distribution in South Florida: a multi-agency co -ordinated monitoring program. In: Porter JW, Porter KG(eds) The Everglades, Florida Bay, and coral reefs of theFlorida Keys: an ecosystem sourcebook. CRC Press, BocaRaton, FL, p 497−522

Green SJ, Akins JL, Maljikovic A, Côté IM (2012a) Invasivelionfish drive Atlantic coral reef fish declines. PLOS ONE7: e32596

Green SJ, Akins JL, Morris JA (2012b) Lionfish dissection: techniques and applications. NOAA Tech Memo NOSNCCOS 139

Hackerott S, Valdivia A, Green SJ, Côté IM, Cox CE, AkinsL, Bruno JF (2013) Native predators do not influenceinvasion success of Pacific lionfish on Caribbean reefs.PLOS ONE 8: e68259

Hill MO (1973) Diversity and evenness: a unifying notationand its consequences. Ecology 54: 427−432

Hill MO (1997) An evenness statistic based on the abun-dance-weighted variance of species proportions. Oikos79: 413−416

Hixon MA, Beets JP (1993) Predation, prey refuges, and thestructure of coral-reef fish assemblages. Ecol Monogr 63: 77−101

Hughes AR, Rooker K, Murdock M, Kimbro DL (2012) Pred-ator cue and prey density interactively influence indirecteffects on basal resources in intertidal oyster reefs. PLOS

ONE 7: e448d39Humann P, Deloach N (2002) Reef fish identification:

Florida, Caribbean, Bahamas. New World Publications,Jacksonville, FL

Johnston MW, Purkis SJ, Dodge RE (2015) MeasuringBahamian lionfish impacts to marine ecological servicesusing habitat equivalency analysis. Mar Biol 162: 2501−2512

Jud ZR, Layman CA (2012) Site fidelity and movement pat-terns of invasive lionfish, Pterois spp., in a Florida estu-ary. J Exp Mar Biol Ecol 414−415: 69−74

Maljkovic A, Van Leeuwen TE, Cove SN (2008) Predation onthe invasive red lionfish, Pterois volitans (Pisces: Scor-paenidae), by native groupers in the Bahamas. CoralReefs 27: 501

McCammon A, Sikkel PC, Nemeth D (2010) Effects of threeCaribbean cleaner shrimps on ectoparasitic mono -geneans in a semi-natural environment. Coral Reefs 29: 419−426

McCune B, Grace JB (2002) Analysis of ecological communi-ties. MjM Software Design, Gleneden Beach, OR

Moe MA (1969) Biology of the red grouper Epinephelusmorio (Valenciennes) from the eastern Gulf of Mexico.Fla Dep Nat Resour Mar Res Lab Prof Pap Ser 10: 1−5

Morris JA (2009) The biology and ecology of the invasiveIndo-Pacific lionfish. PhD dissertation, North CarolinaState University, Raleigh, NC

Morris JA, Akins JL (2009) Feeding ecology of invasive lionfish (Pterois volitans) in the Bahamian archipelago.Environ Biol Fishes 86: 389−398

Morris JA, Akins JL, Barse A, Cerino D and others (2009)Biology and ecology of the invasive lionfishes, Pteroismiles and Pterois volitans. Proc Gulf Caribb Fish Inst 61: 409−414

Mumby PJ, Harborne AR, Brumbaugh DR (2011) Grouper asa natural biocontrol of invasive lionfish. PLOS ONE 6: e21510

Muñoz RC, Currin CA, Whitfield PE (2011) Diet of invasivelionfish on hard bottom reefs of the Southeast USA: insights from stomach contents and stable isotopes. MarEcol Prog Ser 432: 181−193

Oksanen J, Blanchet FG, Kindt R, Legendre P, O’Hara RB,Simpson GL, Wagner H (2011) vegan: community eco -logy package. R package version 1-17. https://CRAN.R-project.org/package=vegan

Okuyama T, Bolker BM (2007) On qualitative measures ofindirect interactions. Ecol Lett 10: 264−271

Ory NC, Thiel M (2013) Host-use patterns and factors influ-encing the choice between anemone and urchin hosts bya caridean shrimp. J Exp Mar Biol Ecol 449: 85−92

Paine RT (1992) Food-web analysis through field measure-ment of per capita interaction strength. Nature 355: 73−75

Pikitch EK, Santora C, Babcock EA, Bakun A and others(2004) Ecosystem-based fishery management. Science305: 346−347

Potts JC, Berrane D, Morris JA (2011) Age and growth oflionfish from the Western North Atlantic. Proc GulfCaribb Fish Inst 61: 314

Preisser EL, Bolnick DI, Benard MF (2005) Scared to death?The effects of intimidation and consumption in preda-tor−prey interactions. Ecology 86: 501−509

Pusack TJ (2013) Coral-reef fishes: insights into larval dis-persal and invasion ecology. PhD dissertation, OregonState University, Corvallis, OR

278

Page 13: Native grouper indirectly ameliorates the negative effects ......juvenile coral reef fishes mediated by the behaviors of the lionfish Pterois volitans and P. miles, piscivores native

Ellis & Faletti: Grouper versus lionfish

R Core Team (2014) R: a language and environment for sta-tistical computing. R Foundation for Statistical Comput-ing, Vienna. www.R-project.org

Raymond WW, Albins MA, Pusack TJ (2015) Competitiveinteractions for shelter between invasive Pacific red lion-fish and native Nassau grouper. Environ Biol Fishes 98: 57−65

Ruttenberg BI, Schofield PJ, Akins JL, Acosta A and others(2012) Rapid invasion of Indo-Pacific lionfishes (Pteroisvolitans and Pterois miles) in the Florida Keys, USA: evi-dence from multiple pre- and post-invasion data sets.Bull Mar Sci 88: 1051−1059

Sax DF, Brown JH (2000) The paradox of invasion. Glob EcolBiogeogr 9: 363−371

Schofield PJ (2009) Geographic extent and chronology ofthe invasion of non-native lionfish (Pterois volitans [Lin-naeus 1758] and P. miles [Bennett 1828]) in the WesternNorth Atlantic and Caribbean Sea. Aquat Invasions 4: 473−479

Shulman MJ, Ogden JC (1987) What controls tropical reeffish populations: recruitment or mortality? An example inthe Caribbean reef fish Haemulon flavolineatum. MarEcol Prog Ser 39: 233−242

Silbiger NJ, Childress MJ (2008) Interspecific variation inanemone shrimp distribution and host selection in theFlorida Keys (USA): implications for marine conserva-tion. Bull Mar Sci 83: 329−345

Stallings CD (2008) Indirect effects of an exploited predatoron recruitment of coral-reef fishes. Ecology 89: 2090−2095

Stallings CD (2009) Predator identity and recruitment ofcoral-reef fishes: indirect effects of fishing. Mar EcolProg Ser 383: 251−259

Strauss SY (1991) Indirect effects in community ecology: their definition, study and importance. Trends Ecol Evol6: 206−210

Trussell GC, Ewanchuk PJ, Matassa CM (2006) Habitateffects on the relative importance of trait- and density-mediated indirect interactions. Ecol Lett 9: 1245−1252

Valdez-Moreno M, Quintal-Lizama C, Gomez-Lozano R,Garcia-Rivas MC (2012) Monitoring an alien invasion: DNA barcoding and the identification of lionfish andtheir prey on coral reefs of the Mexican Caribbean. PLOSONE 7: e36636

Valdivia A, Bruno JF, Cox CE, Hackerott S, Green SJ (2014)Re-examining the relationship between invasive lionfishand native grouper in the Caribbean. PeerJ 2:e348

Venables WN, Ripley BD (2002) Modern applied statisticswith S, 4th edn. Springer, New York, NY

Weaver DC (1996) Feeding ecology and ecomorphology ofthree sea basses (pisces: Serranidae) in the Gulf of Mex-ico. MS thesis, University of Florida, Gainesville, FL

Werner EE, Peacor SD (2003) A review of trait-mediatedindirect interactions in ecological communities. Ecology84:1083–1100

279

Editorial responsibility: Charles Birkeland, Honolulu, Hawaii, USA

Submitted: January 25, 2016; Accepted: June 17, 2016Proofs received from author(s): August 15, 2016

➤➤

➤➤

➤➤

➤➤